Deep learning (DL) is a branch of machine learning that has reshaped the field of cytology. Convolutional neural network (CNN) is a part of deep learning which has the ability to extract features without any external help. The whole scanning images (WSI) contain huge data and it can be handled only by CNN. The DL system mainly contains a feedforward neural network, convolutional neural network, recurrent neural network, autoencoders, generative adversarial networks and a transformer. There are certain limitations of DL. It needs a huge amount of data for proper training. The training process in DL is relatively slow. More powerful computational processing is needed to implement DL. Moreover, the data storage may be another problem in DL because WSI may take huge computer space. The present chapter discusses the basic principles of different DL systems and their potential applications.